The applicants proposed to build a small, low-cost, single slice positron emission tomography (PET) scanner which can be placed inside a conventional magnetic resonance imaging (MRI) system to permit simultaneous acquisition of PET images with MRI, functional MRI (fMRI) or magnetic resonance spectroscopy (MRS) data. The MR compatible PET detector ring will reside inside the MRI magnet and will consist of small lutetium oxyorthosilicate (LSO) scintillator elements which will be read out using multi-clad optical fibers. These optical fibers will transfer the scintillation light to multi-channel photomultiplier tubes and readout electronics which will reside in a radiofrequency (RF) shielded enclosure several meters from the PET detector ring. The data practical difficulties of putting photon detectors and their associated electronics directly inside the MR magnet, and with careful RF shielding eliminates any artifacts in the MR images due to the PET electronics. The system will incorporate a fiducial ring filled with MR contrast agent to allow precise location of the PET imaging slice and direct registration with the MR images. This MR compatible PET scanner will be relatively low in cost, easy to set up and adaptable to a wide range of MR systems. This type of system has many interesting and fundamental applications. It can be used to acquire functional and anatomical information simultaneously, with negligible registration error, allowing the MR images to be used for region of interest definition and possibly partial volume correction of the PET studies. In brain mapping research, it can be used to directly compare fMRI studies with 15O-water PET studies using exactly the same stimulus and imaging environment. It will also permit direct temporal correlation of PET data with MRS, providing additional information in studies of complex metabolic systems. The prototype system outlined in this application would be large enough to image the brains of most non-human primates. If successful, it is possible that the design could ultimately be extended to human brain imaging.